Silicon dioxide nanoparticles decorated on graphene oxide nanosheets and their application in poly(l-lactic acid) scaffold

Graphical abstract

Abstract

Introduction

The aggregation of graphene oxide (GO) is considered as main challenge, although GO possesses excellent mechanical properties which arouses widespread attention as reinforcement for polymers.

Objectives

In this study, silicon dioxide (SiO₂) nanoparticles were decorated onto surface of GO nanosheets through in situ growth method for promoting dispersion of GO in poly(l-lactic acid) (PLLA) bone scaffold.

Methods

Hydroxyl and carboxyl functional groups of GO provided sites for SiO₂ nucleation, and SiO₂ grew with hydrolysis and polycondensation of tetraethyl orthosilicate (TEOS) and finally formed nanoparticles onto surface of GO with covalent bonds. Then, the GO@ SiO₂ nanocomposite was blended with PLLA for the fabrication of bone scaffold by selective laser sintering (SLS).

Result

The results indicated that the obtained SiO₂ were distributed relatively uniformly on surface of GO under TEOS concentration of 0.10 mol/L (GO@SiO₂-10), and the covering of SiO₂ on GO could increase interlayer distance of GO nanosheets from 0.799 nm to 0.894 nm, thus reducing van der Waals forces between GO nanosheets and facilitating the dispersion. Tensile and compressive strength of scaffold containing GO@SiO₂ hybrids were significantly enhanced, especially for the scaffold containing GO@SiO₂-10 hybrids with enhancement of 30.95 % in tensile strength and 66.33 % in compressive strength compared with the scaffold containing GO. Additionally, cell adhesion and fluorescence experiments demonstrated excellent cytocompatibility of the scaffold.

Conclusions

The good dispersion of GO@SiO₂ enhances the mechanical properties and cytocompatibility of scaffold, making it a potential candidate for bone tissue engineering applications.

Introduction

Graphene oxide (GO), with high specific surface area, possesses excellent mechanical properties including superior Young’s modulus and tensile strength [1], [2], [3], thereby arousing widespread attention as reinforcement for polymers. For example, Saha et al. [4] incorporated GO into polyacrylicester polymer via solution blending method and found that Young's modulus of the polymer was improved by 220 % for 4 phr loading of GO. Besides, it has been proved that GO has become an ideal candidate for bone tissue engineering because of its outstanding biocompatibility, which is conducive to cell adhesion and proliferation [5], [6], [7]. Unfortunately, GO is prone to aggregate in the polymer matrix as a result of strong van der Waals force and π-π stacking force among nanosheets, decreasing the contact area of GO with matrix thus failing to exert the advantages of GO on enhancing the properties [8], [9].

It is an efficient strategy to mitigate the agglomeration of GO in the matrix by the addition of another nanomaterial, where the introduced nanomaterial is intercalated into GO interlayers, expanding the lamella space while weakening the interaction force among GO nanosheets [10]. Various approaches to introduce another nanomaterial into GO nanosheets, including physical mixing, non-covalent modification and chemical modification have been employed. Zuo et al. [11] constructed a synergistic dispersion nanosystem of GO and montmorillonite by physical interaction. Niu et al. [12] used physically assembled GO/carbon nanotube hybrids based on the van der Waals forces interactions to hinder the stacking of GO nanosheets. Suprakash et al. [13] obtained a heterostructure resulting from covalent anchoring between hexagonal boron nitride and GO, where hexagonal boron nitride intercalated into the interlayers of GO facilitating their dispersion in matrix. However, the interface bonding between GO and another nanomaterial is weak under physical or non-covalent interaction. While the covalent binding between the two nanomaterials with the help of surface modifiers often leads to the use of chemical reagents that are toxic to the human body through a complex preparation process.

In situ growth is a bottom-up method of direct nucleation and growth of a nanomaterial on another low-dimensional material, thus forming stable chemical bonds and promoting their dispersion [14], [15], [16], [17]. The method is highly reputed owing to the simple preparation process, outstanding compatibility and high bonding strength between the two different nanomaterials. GO nanosheets with a large surface area can support a number of nanomaterials, moreover, they have numerous functional groups, which exhibit great potential on providing numerous nucleation sites for in-situ growth of another nanomaterial [18], [19], [20]. For example, Tayyebi et al. [18] developed a ZnO quantum dots-graphene composite, in which the dangling bonds or functional groups of GO act as nucleation sites to combine quantum dots. Tetraethyl orthosilicate (TEOS) is a precursor of silicon dioxide (SiO₂) and its hydrolysate, namely orthosilicic acid, has abundant hydroxyl groups, which can be absorbed onto the surface of GO through electrostatic interaction [21], [22]. Previous studies had demonstrated that SiO₂ nanoparticles could be decorated on the surface of nanomaterial, through the interaction of polycondensation between hydroxyl of orthosilicic acid and oxygen-containing groups of the nanomaterial as well as the self-polycondensation of orthosilicic acid [23]. Based on the above analysis, the in situ growth of SiO₂ nanoparticles onto the surface of GO nanosheets via hydrolysis and condensation of TEOS, might be a feasible way to prevent the aggregation of GO nanosheets in matrix. And this might effectively exert the advantages of GO on enhancing the properties of the polymer.

In this study, in situ growth method was adopted for decorating SiO₂ nanoparticles onto the surface of GO nanosheets through hydrolysis and condensation of TEOS in order to promote the dispersion of GO in biopolymer poly(l-lactic acid) (PLLA) bone scaffold prepared via selective laser sintering (SLS) [24], [25], [26], [27]. PLLA [28], [29], [30] was used as bone scaffold material due to its outstanding processability, biocompatibility and biodegradability, while the insufficient mechanical strength restricted its wide range of application. It is speculated that the SiO₂ decorated GO (GO@SiO₂) might be well dispersed in PLLA matrix and play a positive effect on enhancing mechanical properties of scaffold, making it a potential candidate for bone tissue engineering applications. The physicochemical properties of GO@SiO₂ hybrids were characterized by X-ray diffraction (XRD), Fourier transform infrared spectroscopy (FTIR) and X-ray photoelectron spectroscopy (XPS). The surface morphologies of GO@SiO₂ hybrids as well as the dispersion states in PLLA bone scaffold were observed by scanning electron microscopy (SEM). The tensile and compressive testings were carried out to estimate mechanical properties of the scaffold. Additionally, cell adhesion and fluorescence experiments were conducted for assessing cytocompatibility.

Materials and methods

Materials

GO nanosheets (purity > 98 %) with diameters in the range of 8 ∼ 15 μm were supplied by Chengdu Organic Chemicals Co., ltd., Chinese Academy of Sciences (Chengdu, China). TEOS (purity > 99 %) was procured from Sinopharm Chemical Reagent Co., ltd. (Shanghai, China). PLLA powders (inherent viscosity: 2.0 dL/g, molecular weight: 150 kDa) were purchased from Shenzhen Polymtek Biomaterial Co., ltd. (Shenzhen, China). Acetic acid (C₂H₄O₂), ethanol (C₂H₅OH, purity > 98 %), sodium hydroxide (NaOH), distilled water (PH0629), phosphate buffered solution (PBS) and simulated body fluid (SBF) were procured from Sinopharm Chemical Reagent Co., ltd. (Beijing, China).

Preparation of GO@SiO₂ hybrids

GO@SiO₂ hybrids were fabricated via in-situ hydrolysis and condensation of TEOS, and the whole process is displayed in Fig. 1 (a). Concretely, the appropriate amount of TEOS was added into the alcohol-water (4:1 v/v) solution to prepare the silane solution (100 mL) with TEOS concentrations of 5.0 × 10^-2, 1 × 10^-1 and 1.5 × 10^-1 mol/L, respectively. The pH of silane solution was controlled at 4 by adding acetic acid which might promote the hydrolysis of TEOS into orthosilicic acid [31], and then stirred at 20 rpm by a magnetic stirrer. Meanwhile, GO nanosheets (25 mg) were added into alcohol-water (25 mL, 4:1 v/v) solution with sonication with the purpose of getting homogeneous dispersion of GO. Then the GO solution (25 mL) was mixed with the silane solution (100 mL) and stirred at 50 rpm. Subsequently, the pH of mixture was controlled at 9 by adding NaOH solution which might promote the condensation reaction thus facilitating the growth of SiO₂ nanoparticles, and then the solution was stirred at 50 rpm for 1 h at 65 ℃. In this process, GO and orthosilicic acid might be attracted to each other by electrostatic interaction among functional groups, and the polycondensation occurred between the silicone alcohol group and the oxygen-containing of GO, with NaOH as catalysis [32]. The mixed solution was centrifuged at 8000 rpm and washed by ethanol, and subjected to drying in an oven at 50 ℃ for 24 h. The GO@SiO₂ hybrids obtained with TEOS concentrations of 5.0 × 10^-2, 1 × 10^-1 and 1.5 × 10^-1 mol/L were denoted as GO@SiO₂-5, GO@SiO₂-10 and GO@SiO₂-15, respectively.

Fig. 1.

(a) The synthesis process of GO@SiO₂ hybrids. (b) The schematic of the interaction between GO and TEOS, (b1) the schematic representation of TEOS bonding reaction with GO, (b2) the SEM micrograph of GO@SiO₂ hybrids.

Preparation of composite powders and scaffolds

The PLLA-based composite powders containing 1.5 wt% of GO nanosheets or GO@SiO₂ hybrids were prepared using the solution blending method. In a typical synthesis of PLLA-based composite powders containing GO (denoted as PG), 0.045 g of GO nanosheets were firstly dispersed in 50 mL ethanol by sonication. The prepared solution was blended with 2.955 g PLLA powders, followed by stirring at 20 rpm to get an even suspension. Subsequently, the composite powder was eventually obtained after centrifuging and drying. The fabrication procedures of the PLLA-based composite powders containing GO@SiO₂-5, GO@SiO₂-10, GO@SiO₂-15 were similar to that of PG, and the obtained composite powders were denoted as PG5, PG10 and PG15, respectively.

The fabrication of three-dimensional scaffolds was achieved by SLS system. Before sintering, a design 3D scaffold model was converted into STL format and imported into SLS system then the slice data of scaffold was obtained. The preparation of PLLA, PG, PG5, PG10 and PG15 scaffolds was under the same processing parameters with a spot diameter of 50 μm, laser power of 2 W, and scanning speed of 80 mm/s. When preparing for sintering, a detailed preparation process with the following steps is revealed in Fig. 4(b): (Ⅰ) the powder with thickness between 0.1 and 0.2 mm was distributed uniformly on the sintering platform by the roller, (Ⅱ) one powder layer was selectively sintered by the laser beam according to the slicing date of scaffold model, (III) the platform moved vertically a corresponding height of one powder layer followed by a new powder layer spreading, and then the new powder layer was sintered. Finally, a scaffold was obtained after repeating the layer by layer sintering process.

Fig. 4.

(a) Digital and SEM images of PLLA, PG and PG10 composite powders. (b) Schematic illustration of the SLS procession. (c) Digital images of PLLA and PG10 scaffolds.

Characterization

FTIR (Thermo Scientific Nicolet iS5, Massachusetts, USA) was utilized functional groups of GO before and after introduction of TEOS. XPS (Thermo Scientific K-Alpha, Massachusetts, USA) was employed to examine phase composition of GO@SiO₂ hybrids. XRD (Rigaku Ultima IV, Tokyo, Japan) was performed at a rate of 5°/min from 5° to 90° to characterize the chemical components of GO with and without the formation of SiO₂ nanoparticles. Raman (Horiba LabRAM HR Evolution, Japan) was conducted for analyzing the structure of GO and GO@SiO₂ hybrids. SEM (Phenom-World BV, Eindhoven, Netherlands) was employed in investigating surface morphologies and dispersion states of GO@SiO₂ hybrids as well as the corresponding composite scaffolds, provided with energy-dispersive spectroscopy (EDS) investigating the elemental composition. Before SEM observation, the samples were subjected to gold sputtering treatment for 120 s.

The mechanical behaviors of scaffolds were examined by a universal testing machine (YAW-300C, Shandong, China) in terms of tensile and compressive testings, utilizing dumbbell-like and cylindrical samples with the dimension of 20 mm × 4 mm × 2 mm and Φ5 mm × 10 mm, respectively. The stress–strain curves were recorded under a set crosshead speed of 0.5 mm/min. Thermogravimetric analysis (TGA, Dazhan STA-200, Dazhan Electromechanical Technology Institute, Nanjing, China) was recorded in the range of 50–500 ℃ to study the thermal stability of composite powders. The water contact angle (WAC, Rame Hart 500-F1, RameHart Instrument Co., USA) was measured to evaluate the hydrophilicity of scaffolds. At least five valid samples were tested for each group.

Swelling behavior and degradability

The swelling behaviors of scaffolds were measured by the gravimetric method as reported [33]. The samples with a dimension of 10 mm × 10 mm × 2 mm were weighed (denoted as W₀) and immersed in PBS at 37 ℃. After soaking for a predetermined time, the samples were fetched out and wiped off with absorbent paper. Then the weights of wet samples after being immersed for a certain time (t) were measured (denoted as W_1t) to calculate the swelling ratios (S_t, %) according to the following equation [34]:

$${\mathrm{S}}_{\mathrm{t}}\left(\%\right)=\frac{{\mathrm{W}}_{1\mathrm{t}}-{\mathrm{W}}_0}{{\mathrm{W}}_0}\times 100\%$$

The degradation properties of the scaffolds were investigated on the basis of weight loss and surface morphology evolution after immersion in PBS at 37 ℃ for 7, 14 and 21 d. The samples were taken out at a predetermined time and subjected to cleaning and drying, and the surface morphologies were characterized by SEM. The residual weight of samples after dried was measured (denoted as W_2t). And the residual weight fraction of samples (R_t, %) after being immersed for a certain time was calculated by the following relation [35]:

$${\mathrm{R}}_{\mathrm{t}}\left(\%\right)=\frac{{\mathrm{W}}_{2\mathrm{t}}}{{\mathrm{W}}_0}\times 100\%$$

The degradation rate (D_t, %/d) could be calculated with the following formula [36]:

$${\mathrm{D}}_{\mathrm{t}}\left(\%/\mathrm{d}\right)=\frac{{\mathrm{R}}_0-{\mathrm{R}}_{\mathrm{t}}}{\mathrm{t}}$$

Biomineralization

The biomineralization behavior of the scaffolds was characterized on the basic of the ability of apatite formation by immersing in SBF. After immersing for 14 d, the scaffolds were cleaned with distilled water, then subjected to drying, and finally investigated by SEM and EDS for observing the apatite formation of the scaffolds. The tensile testing was carried out utilizing the dumbbell-like samples after immersion in SBF for 14 d.

Cytocompatibility

Cellular compatibility of PLLA, PG and PG10 scaffolds was investigated on the basis of cell adhesion and fluorescence experiments. Human osteoblast-like cells (MG63, Manassas, Virginia, USA) were incubated in a 24 well-plate containing Dulbecco's modified Eagle's medium (DMEM) supplemented with 10 % fetal bovine serum and 1 % antibiotic–antimycotic solution in 5 % CO₂ at 37 °C. Subsequently, the cells were seeded in each sample under a concentration of 2.0 × 10⁵ cells/cm² in culture plate containing the DMEM and cultured in a humidified incubator at 37 ℃ with 5 % CO₂ for 1, 3 and 5 d SEM was conducted to investigate the cell adhesion after the samples were treated with the glutaraldehyde and PBS, and the PhotoShop software was employed to calculate the cell area. Cells were stained with calcein acetoxymethyl (calcein AM) as well as propidium iodide (PI) and then investigated via fluorescence microscopy (Olympus Corporation, Tokyo, Japan). In the stained images, live cells exhibited green fluorescence while dead cells showed red fluorescence. And cell density could be obtained via calculating ratio of the live cells’ number to image’s area.

Statistical analysis

Data from all quantitative experiments were reported as mean ± standard deviation (SD). The Statistical Product and Service Solutions (SPSS) software was employed to calculate the statistical significance. p < 0.05 (*) was considered statistically significant and p < 0.01 (**) was considered very significant.

Results and discussion

Characteristic of GO@SiO₂ hybrids

The schematic of interaction between GO and TEOS is displayed in Fig. 1 (b). TEOS was hydrolyzed to orthosilicic acid and the hydrolysis rate of TEOS could be promoted in the presence of acetic acild. Then orthosilicic acid could be adsorbed onto GO via electrostatic interaction between the hydroxyl groups of orthosilicic acid and the hydroxyl as well as the carboxylic groups of GO, and then formed stable covalent bonds with GO via polycondensation interaction. Besides, the self-polycondensation of orthosilicic acid occurred at the same time and the rate could be promoted using NaOH as catalysis. Finally, SiO₂ nanoparticles were formed on the surface of GO nanosheets, as shown in Fig. 1 (b2).

The FTIR spectra of GO along with corresponding hybrids are displayed in Fig. 2 (a) for analyzing functional groups. For GO, peak at 3380 cm⁻¹ was a typical feature of –OH vibration of hydroxyl groups [37], and the absorption peaks at 1734 cm⁻¹ and 1627 cm⁻¹ were ascribed to the C O vibration of carboxylic groups and C C skeletal vibrations in phenol ring, respectively [38]. These peaks were typical characteristic peaks in GO marked by light blue. After being treated by TEOS at different concentrations, the intensity of peak at 1734 cm⁻¹ decreased obviously, which might be attributed to the conversion of carbonyl groups to the Si—O—C band. Besides, new peaks marked by light yellow were observed. The new peak at 1189 cm⁻¹ belonged to Si—O—C asymmetric vibration, indicating polycondensation between hydroxyl groups on orthosilicic acid and oxygen-containing groups on GO [39]. The peaks at 1096 cm^-1and 797 cm⁻¹ were related to the Si-O-Si symmetric vibration [40], and the peak at 468 cm⁻¹ was related to the Si-O-Si bending vibration [41], indicating the self-polycondensation of orthosilicic acid. It could conclude from the analysis of FTIR that SiO₂ nanoparticles were successfully introduced onto the GO surface through the hydrolysis and condensation of TEOS.

Fig. 2.

The characterization of GO, GO@SiO₂-5, GO@SiO₂-10 and GO@SiO₂-15. (a) FTIR spectra. XPS investigation: (b) wide-scan survey; (c) The partial enlargement of (b); (d) C1s spectra of GO; (e) C1s spectra of GO@SiO₂-10; (f) O1s spectra of GO; (g) O1s spectra of GO@SiO₂-10; (h) Si2s spectra of GO@SiO₂-10. It was worth noting that the red line (denoted the fitted data) was highly overlapped with the black line (represented the original data) in the narrowed spectra. (i) XRD patterns; (j) The partial enlargement of (h). (k) Raman spectra; (l) The partial enlargement of (k).

XPS was employed in further characterizing chemical compositions of GO and GO@SiO₂ hybrids, and the results are shown in Fig. 2 (b-h). As shown in Fig. 2 (b, c), the C1s and O1s signals were observed in the wide-scan survey XPS spectra of all GO and GO@SiO₂ hybrids, however, for GO@SiO₂ hybrids, new signals of Si2p and Si2s were observed, which might be ascribed to the formation of SiO₂ nanoparticles onto GO surface [42]. As shown in Table 1, for GO@SiO₂-10, the contents of C, O and Si elements were changed from 70.15 % to 13.09 %, 29.85 % to 66.31 % and 0 % to 20.60 %, respectively, compared with GO. The change in the element contents was due to the formation of SiO₂ nanoparticles. Additionally, the content of Si element changed among GO@SiO₂ hybrids, which might be attributed to the fact that more SiO₂ nanoparticles were grown onto the surface of GO with increasing concentration of TEOS. The narrowed spectra of GO and GO@SiO₂-10 are shown in Fig. 2 (d-h). For GO, the C1s spectra of GO could be separated into four peaks ascribed to C—C (284.8 eV), C—O—C (286.7 eV), C O (287.6 eV) and O-C O (289 eV) shown in Fig. 2 (c). There exhibited two more new peaks related to Si—O—C (285.7 eV) and C-Si (283.8 eV) in the C1s spectra of GO@SiO₂-10 compared with GO, as demonstrated in Fig. 2 (d) [43]. As shown in Fig. 2 (f), the O1s spectra of GO included four peaks attributed to O-C O (532.1 eV), C—O (531.0 eV), C-OH (532.6 eV) and C—O—C (534.1 eV). As shown in Fig. 2 (g), for GO@SiO₂-10, a new peak corresponded to Si—O—C (533.6 eV) was fitted and the peak at 532.7 eV could be assigned to Si-O-Si [44]. According to Fig. 2 (h), the Si spectra could be fitted by three peaks assigned to Si-OH (102.9 eV), Si-O-Si (103.4 eV) and Si—O—C (103.6 eV) [45]. These observations validated the in situ growth of SiO₂ onto GO surface, which was in accordance with the analysis of FTIR.

Table 1.

The elementary composition of GO and GO@SiO₂-5, GO@SiO₂-10 and GO@SiO₂-10 hybrids obtained from XPS.

Samples	C (at%)	O (at%)	Si (at%)
GO	70.15	29.85	–
GO@SiO2-5	65.89	30.44	3.67
GO@SiO2-10	13.09	66.31	20.60
GO@SiO2-10	10.55	61.90	27.55

XRD was performed for investigating phase structure of GO as well as GO@SiO₂, and the interlayer distance could be calculated with the obtained X-ray wavelength (λ = 1.5418 Å) as well as the angle between the surface of crystal and incident rays (θ) according to Bragg’s equation: nλ = 2dsinθ [46]. The GO exhibited a peak centered at 2θ = 11.06° which corresponded to 001 planes [47], which indicated the highly ordered structure of GO nanosheets and the interlayer spacing values were 0.799 nm according to Bragg’s equation. After being treated by TEOS at different concentrations, the peak attributed to silica nanoparticles arose at around 2θ = 23.40° [48], besides, the peak centered at 2θ = 11.06° showed varying degrees of blue shift, and the blue shift increased with increasing concentration of TEOS. In detail, for GO@SiO₂-5, GO@SiO₂-10 and GO@SiO₂-15, there were characteristic peaks at 10.80°, 9.88° and 9.58°, respectively, and the interlayer spacing values were 0.818 nm, 0.894 nm and 0.922 nm, respectively. This demonstrated that the covering of SiO₂ on GO could increase the interlayer distance of GO nanosheets, and the increase is most pronounced when GO was treated with the highest concentration of TEOS.

The Raman spectra of GO and GO@SiO₂ hybrids are shown in Fig. 2 (k, l). Typically, the GO exhibited a strong D peak at 1356 cm⁻¹ refer to the sp² ordered crystalline graphite-like structure and a strong G peak at 1588 cm⁻¹ refer to the sp³ disordered carbon structure. For GO@SiO₂ hybrids, the G peak was shifted towards higher wave number (1589 cm⁻¹, 1591 cm⁻¹, 1592 cm⁻¹ for GO@SiO₂-5, GO@SiO₂-10, GO@SiO₂-15, respectively). Moreover, the peak intensity ratio (I_D/I_G) increased from 0.92 for GO to 0.94, 0.96, 0.97 for GO@SiO₂-5, GO@SiO₂-10, GO@SiO₂-15, respectively, indicating improved disorder level of the layers for GO@SiO₂ hybrids, which might lead to the expanded layer spacing as previous reported [49], [50], [51].

SEM and EDS were conducted to provide intuitive evidence of the surface morphology difference among GO and corresponding hybrids treated by different concentrations of TEOS, and the results are displayed in Fig. 3. GO exhibited a typical structure with stacked sheets and a relatively smooth surface with some wrinkles as illustrated in Fig. 3 (a1, a2) [52]. For GO@SiO₂-5, some white spherical silica nanoparticles were uniformly distributed on the surface of GO as shown in Fig. 3 (b1, b2), and the EDS profiles also confirmed the presence of Si element compared with the pristine GO. For GO@SiO₂-10, there existed more silica particles, and the corresponding elemental mapping images manifested that there was a slight agglomeration of SiO₂ nanoparticles as shown in Fig. 3 (c1, c2). And SiO₂ nanoparticles uniformly distributed on the surface of GO could increase the contact area thus facilitating the dispersion of GO in the polymer matrix [53]. For GO@SiO₂-15, a large amount of SiO₂ nanoparticles were stacked on the surface of GO, as shown in Fig. 3 (d1, d2). And it was clearly observed from the elemental mappings that the distribution of Si was highly concentrated indicating the agglomeration of SiO₂ nanoparticles. Combing the results of FTIR, XPS, XRD and SEM-EDS, it could be concluded that the SiO₂ nanoparticles were relatively uniformly grown onto the surface of GO nanosheets via covalent binding with TEOS concentrations of 0.1 mol/L.

Fig. 3.

Surface morphologies of (a1) GO, (b1) GO@SiO₂-5, (c1) GO@SiO₂-10 and (d1) GO@SiO₂-15 hybrids. (a2-d2) The partial enlargement images of (a1-d1) with the corresponding elemental mapping images and energy dispersive X-ray spectra.

Characteristic of composite powders and scaffolds

The morphologies of PLLA, PG and PG10 powders are shown in Fig. 4 (a). PLLA powders presented white and exhibited an irregular shape with approximately 10 ∼ 120 μm in size.

PG and PG10 powders presented grey and light brown, respectively, and the GO and GO@SiO₂ hybrids were distributed uniformly on the surface of PLLA particles after the solution-blending. The PLLA and PG10 scaffolds were prepared by SLS, a typical additive manufacturing technology that could realize the preparation of personalized shape and controllable internal porous structure of bone scaffold [54], [55]. The digital images of PLLA and PG10 scaffolds with a color change from white to black are shown in Fig. 4 (b, c). The well-designed porous scaffolds presented as a hexagonal cylinder with a dimension of Φ16 mm × h10 mm combined with the interpenetrating pores in the size of about 600 μm, which was believed to provide a suitable microenvironment for nutrients conveying and cell proliferating [56], [57], [58].

As shown in Fig. 5 (a-h), SEM was performed with the purpose of investigating the dispersion states of GO and GO@SiO₂ hybrids in PLLA matrix, and corresponding schematic diagrams are shown in Fig. 5 (i-l). For the PG scaffold, there was a moderate agglomeration as shown in Fig. 5 (a, e and i). GO nanosheets were prone to aggregate in PLLA matrix due to the strong van der Waals force and π-π stacking force, which might decrease the contact area of GO with polymer matrix and even lead to defects in the scaffold [59]. Comparatively, the dispersion state of GO@SiO₂-5 in PLLA matrix was improved to a certain extent despite the observation of a small agglomeration as shown in Fig. 5 (b, f and j). And GO@SiO₂ hybrids exhibited a relatively rough surface in PLLA matrix compared with GO whose surface presented a smooth condition as shown in Fig. 5 (i, j), which was in accordance with the result of surface morphologies discussed previously (Fig. 3). As shown in Fig. 5 (c, g), the PG10 scaffold showed a significant improvement in PG state with no obvious aggregation at the same loading levels of filler. Under an appropriate concentration of TEOS, the obtained SiO₂ nanoparticles were distributed relatively uniformly on the surface of GO, playing an important role in the formation of synergistic system for promoting dispersion. In detail, the obtained SiO₂ nanoparticles acted as steric hindrance between GO nanosheets and facilitating the dispersion. Therefore, the advantage of GO on enhancing the properties of the polymer might be exerted more effectively [60]. However, for the PG15 scaffold, a slight aggregation was observed for GO@SiO₂-15 in PLLA, which might be due to the excessive SiO₂ nanoparticles which were easily formed agglomerates.

Fig. 5.

SEM surface morphology of (a) PG, (b) PG5, (c) PG10 and (d) PG15 scaffolds. (e-h) The partial enlargement images of (a-d). (i-l) The corresponding schematic diagrams of fillers dispersed in the PLLA matrix.

Properties of scaffolds

The effect of GO and GO@SiO₂ hybrids on the mechanical properties of scaffolds was assessed via tensile combined with compressive testings. The tensile stress–strain curves along with tensile strength and modulus of PLLA, PG, PG5, PG10 and PG15 scaffolds are exhibited in Fig. 6 (a, b). For the PG scaffold, the tensile strength and modulus increased from 6.3 MPa and 509.5 MPa for the PLLA scaffold to 8.4 MPa and 1186.8 MPa, respectively. This could be attributed to the reinforcing effects of GO nanosheets. After the formation of SiO₂ nanoparticles onto the surface of GO nanosheets, the tensile strength and modulus of the PG5 and PG10 scaffolds increased remarkably, especially for the PG10 scaffold, possessing the highest tensile strength as well as modulus of 11.0 MPa and 1700.5 MPa with the enhancement of 30.95 % and 43.28 %, respectively, comparing to the PG scaffold. This might be attributed to the introduction of SiO₂ nanoparticles, which not only facilitated the dispersion of GO in PLLA matrix for a higher load transfer capability thereby taking full advantage of GO on enhancing the properties of the PLLA, but combine the advantage of GO and SiO₂ to improve the mechanical properties of scaffolds. However, the tensile strength and modulus of the PG15 scaffold were 10.3 MPa and 1389.6 MPa, respectively, exhibited a slight decrease at the same loading (1.5 wt%) compared to the PG10 scaffold. The compressive strength and modulus determined from the compress stress–strain curves followed similar trends to tensile properties, as shown in Fig. 6 (c, d). And the PG10 scaffold possessed maximum compressive strength and modulus of 16.3 MPa and 164.8 MPa with the enhancement of 66.33 % and 114.30 %, respectively, compared with the PG scaffold. In conclusion, the PG10 scaffold was considered optimal in terms of mechanical properties.

Fig. 6.

(a) Tensile stress–strain curves and (b) tensile strength and modulus of scaffolds. (c) Compressive stress–strain curves and (d) compressive strength and modulus of scaffolds. (e) TGA curves of scaffolds. (f) The water contact angle of scaffolds. (The error bar represents the standard deviation (SD). * represents p < 0.05 compared with PLLA scaffold).

Based on the analysis of tensile and compressive testing, it could conclude that the mechanical properties of the PG scaffold were better than the PLLA scaffold, which could be ascribed to reinforcing effects of GO. The mechanical properties of the PG5, PG10 and PG15 scaffolds were higher than that of the PG scaffold, which might attribute to the following factors: for one thing, it is said that the agglomeration of hybrids created local stress concentration zones in the matrix leading to a reduced load transfer capability, which causes declined mechanical properties [61], [62], [63]. The load transfer capability might be increased after the formation of SiO₂ onto GO surface due to the good dispersion state in matrix, which caused the enhancement of mechanical properties for scaffolds. For another thing, the SiO₂ nanoparticles in-situ growing on the surface of GO could increase the contact area between hybrids with matrix, and the SiO₂ nanoparticles could absorbing more crack energy making transfer of stress more efficiently [64], [65], which improved the mechanical properties of scaffolds. However, for the PG15 scaffold, when excessive SiO₂ nanoparticles grew on the surface of GO, it was prone to aggregate thus leading to a decrease in mechanical properties compared with PG10 scaffold.

The thermal stability of scaffolds was determined by TGA, as shown in Fig. 6 (e). Also, the decomposition temperature (T_d, the values of 5 % weight loss temperature) was recorded to assess thermal stability of the scaffold samples [66]. The T_d value of pure PLLA scaffold was 340.0 ℃. Clearly, the incorporation of GO or GO@SiO₂ hybrids was advantageous to enhance the thermal stability of PLLA, and the PG 10 scaffold demonstrated the best thermal stability among other samples, revealing an increment of 11.5 ℃ in T_d values compared with that of PLLA. This might be attributed to the oxygen barrier properties of nanosheets and the homogenous dispersion of GO@SiO₂ hybrid, which acted as a barrier against decomposition [67]. The hydrophilicity was determined from water contact angle testing. The PLLA scaffold exhibited hydrophobicity with angles values of 118.6°. The PG scaffold exhibited a decrease of water contact angle in comparison to the PLLA scaffold, which might be attributed to the addition of the GO with many hydrophilic groups. For the PG5, PG10 and PG15 scaffolds, the water contact angle gradually decreased and exhibited less hydrophobic characteristics than unmodified PLLA with contact angles values of 85.1°, 84.8°, 82.2°, respectively, indicating that the hydrophilic materials of SiO₂ nanoparticle possessed a positive effect on hydrophilicity. In conclusion, the introduction of GO and GO@SiO₂ into the PLLA matrix could improve the hydrophilic of PLLA.

The swelling behavior was investigated to confirm the effect of the aqueous environment on scaffolds' properties. The changing of swelling ratios versus immersing time was demonstrated in Fig. 7 (a). The swelling rations of all scaffolds exhibited an upward trend, and reached the equilibrium swelling ratio after about 9 h. The equilibrium swelling ratio of PLLA, PG, PG5, PG10 and PG15 were 2.09 %, 3.14 %, 3.52 %, 3.72 and 4.06 %, respectively. The results indicated that the addition of GO and GO@SiO₂ could enhance the water storage capacity of PLLA, which might be attributed to the enhancement in hydrophilicity of scaffolds. The curves of residual weight fraction are exhibited in Fig. 7 (b, c). The weight loss of all the scaffolds exhibited a gradual decrease with increasing immersed time. The PLLA scaffold exhibited a small change of residual weight in the whole process indicating poor degradability. Comparatively, the introduction of GO and GO@SiO₂ hybrids induced a more apparent change of residual weight and a faster degradation rate of 0.18 %/d, 0.19 %/d, 0.22 %/d and 0.26 %/d for the PG, PG5, PG10 and PG15 scaffolds after immersing for 21 d, which showed a similar trend with swelling rations. The morphologies evolution of the scaffolds after immersing in PBS for 21 d was shown in Fig. 7 (d1-d5). The PLLA scaffold exhibited a flat surface without hole. While some tiny holes were observed in the PG scaffold and more holes with larger size in the PG15 scaffold, showing typical bulk erosion morphologies. Based on the above analysis, it could be concluded that the degradation of PLLA was accelerated after introducing GO or GO@SiO₂ hybrids, which might be due to the improvement of the hydrophilic. Previous studies have shown that more interfaces in composites were induced by GO, and then provided inflow channels for water molecules, thus accelerating the degradation of PLLA [68]. The presence of SiO₂ could also play a positive effect on facilitating the degradation of PLLA as previously reported. In summary, the introduction of GO and GO@SiO₂ into the PLLA matrix could facilitate the degradation of PLLA [69], [70].

Fig. 7.

(a) The swelling kinetics of scaffolds after immersing in PBS at 37 ℃. (b) The weight loss and (c) degradation rate of scaffolds after immersing in PBS at 37 ℃. SEM surface morphologies of (d1) PLLA, (d2) PG, (d3) PG5, (d4) PG10 and (d5) PG15 scaffolds after immersing in PBS for 21 d. SEM surface morphologies and EDS spot scanning spectra of (e1) PLLA, (e2) PG, (e3) PG5, (e4) PG10 and (e5) PG15 scaffolds after immersing in SBF for 14 d; (f) Tensile stress–strain curves and (g) tensile strength and modulus of scaffolds after immersing in SBF for 14 d. (h) The increasing ratio before and after immersion. (The error bar represents the SD. * represents p < 0.05 compared with PLLA scaffold).

The in vitro biomineralization experiment was conducted to evaluate the bioactivity of the scaffolds, and the results are displayed in Fig. 7 (e1-e5). After immersion in SBF for 14 d, no obvious deposit was found on the PLLA scaffold indicating PLLA had no bioactivity. Comparatively, some apatites with globular shape were observed in the composite scaffolds and the peak of both Ca and P was detected in the EDS spectra. The mineralization levels followed a trend of PG15 > PG10 > PG5 > PG. The apatite could bond the scaffold to living bone and enhance bone regenerative potential [71]. The good ability of apatite formation of the composite scaffolds might contribute to the following factors. For one thing, the oxygenic functional groups on GO could absorb Ca²⁺ via electrostatic interaction and induce mineralization [72], [73]. For another thing, SiO₂ could also provide sites for apatite nucleation and mineralization [74], [75], enhancing the bioactivity of the scaffolds.

The tensile testing was conducted to assess the mechanical properties of the scaffolds after immersion in SBF at 37 ℃ for 14 d, and the results are shown in Fig. 7 (f-h). It was found that the tensile properties of the scaffolds followed a similar trend to that of the scaffolds before SBF immersion, with the PG10 scaffold possessing the highest tensile strength. The degree of tensile enhancement of the PG5, PG10 and PG15 scaffolds compared with the PG scaffold was analyzed by the increasing rate (I, %) before and after immersion. The increasing rate was calculated according to the following equation: I=(S₁-S₂)/S₂ × 100 %, where S₁ represented the tensile strength of the scaffolds containing GO@SiO₂ hybrids and S₂ for the PG scaffold. The results are displayed in Fig. 7 (h). The enhancement effect in tensile strength of GO@SiO₂ hybrids on matrix after immersion was worse than that before immersion. Especially, the tensile strength of the PG15 scaffold after immersion was 41.03 % higher than that of the PG scaffold, while the increasing rate was 62.57 % before immersion. This might associate with the change in the degree of hydrophilicity and biodegradability, which was consistent with previous research [76], [77].

Cytocompatibility

Cell experiment was carried out in order to assess cytocompatibility of the PG10 scaffold, while the PLLA and PG scaffolds was used as control. The SEM images of MG63 cells culturing on the PLLA, PG and PG10 scaffolds for 1, 3 and 5 d are presented in Fig. 8 (a), manifesting that all scaffolds provided a suitable environment for cell adhesion and proliferation [78]. In detail, for 1 d, MG63 cells on the PLLA scaffold were separated from each other with elongated spindle shape, while cells on the PG scaffold were plumper and exhibited some filopodia extensions on the PG10 scaffold. For 3 d, with the growth of cells, filopodia intertwined with each other, and the coverage area of cells increased resulting in a shape-change to irregularly polygonal. For 5 d, there formed a cell layer on scaffolds as a result of the growth and extension of the MG63 cells. The cell-spreading area expanded along with culturing time prolonging on all scaffolds, as displayed in Fig. 8 (b). And PG10 scaffold exhibited the best cytocompatibility in the light of the area among three scaffolds, followed by the PG scaffold, demonstrating that the introduction of GO and SiO₂ into PLLA both could improve cell proliferation.

Fig. 8.

(a) MG63 cell morphologies cultured on the PLLA, PG and PG10 scaffolds for 1, 3 and 5 d, and (b) the corresponding statistical of cell spreading area. (The error bar represents the SD. * represents p < 0.05 compared with PLLA scaffold).

The cell viability of the PG10 scaffold was evaluated by the fluorescence experiment. The fluorescence images of MG63 on the PLLA, PG and PG10 scaffolds with living cells (in green) and dead cells (in red) for a different time are exhibited in Fig. 9 (a), along with corresponding cell densities shown in Fig. 9 (b). Clearly, the living cells spread well with filopodia growing and the number of living cells was proportional to different periods [79], [80], [81], while almost no dead cells were visible in either scaffold. Besides, the PG and PG10 scaffolds had higher levels of cell viability compared to the PLLA scaffold at all time points which was in line with the results of the cell adhesion experiment, especially for the PG10 scaffold, possessing the highest cell viability. This demonstrated that the introduction of GO into PLLA matrix could enhance cell proliferation, and the enhancing effect could be further increased by in-situ growing of SiO₂ onto GO surface. Taken together, the results of cell adhesion and fluorescence experiments indicated that the PG10 scaffold was superior to the PLLA and PG scaffolds in promoting the growth of MG63 cells, which might contribute to the following factors. On the one hand, the improved hydrophilicity of the PG10 scaffold by adding GO@SiO₂ hybrids was also beneficial for cell adhesion and proliferation. On the other hand, GO@SiO₂ hybrids exhibited a large and rough surface which provided more cell adhesion sites. Besides, Si from GO@SiO₂ hybrids played a significant role in stimulating cellular behaviors thus promoting cell proliferation.

Fig. 9.

(a) Fluorescence images of MG63 cells cultured on the PLLA, PG and PG10 scaffold for 1, 3 and 5 d, and (b) the corresponding statistical of cell density. (The error bar represents the SD. *, ** represent p < 0.05 and p < 0.01 compared with PLLA group, respectively).

Overall, PG10 scaffold possessed appropriate physicochemical properties exhibiting great potential for bone tissue engineering application. It is an effective strategy to add nanofillers into biomaterials for enhance the performance of the scaffolds as reported by previous study [82]. In fact, considerations must be placed on the properties of the employed materials, as artificial bone scaffolds require various properties including appropriate mechanical properties and biodegradability as well as excellent biocompatibility to meet clinical needs [83], [84], [85], [86]. Biopolymer, bioceramic and combinations thereof have captured more eyes in promoting bone regeneration [87], [88], [89], [90]. For example, Cui et al. [87] reported a unique composite scaffold with the addition of gene vector into PLGA/HA matrix for accelerating osteogenic expression. Those studies are very useful for the enhancement of the positive interactions between scaffold and bone defect and will become an important strategy for our future work.

Conclusions

The present study reported that a unique hybrid was successfully obtained by SiO₂ nanoparticles in situ decorated onto the surface of GO nanosheets through hydrolysis and condensation of TEOS for improving the dispersion of GO in PLLA. And the scaffolds were fabricated by SLS for preparing a personalized three-dimensional interconnected porous structure. The results demonstrated that SiO₂ nanoparticles grown on the surface of GO played a role in steric hindrance against the aggregation of GO in PLLA scaffold. The tensile and compressive testings showed that the GO@SiO₂ hybrids were more effective in enhancing PLLA mechanical properties and the PG10 scaffold exhibited the highest enhancement, increasing 30.95 % in tensile strength and 66.33 % in compressive strength compared with the PG scaffold. In addition, the incorporation of GO@SiO₂ could compensate the poor degradation property and bioactivity of PLLA to some extent. And the PG10 scaffold also exhibited excellent cytocompatibility and viability compared with PLLA and PG scaffolds. In summary, the PG10 scaffold might open up new possibility for bone tissue engineering. In the future, an important strategy for our work is conducting in vivo studies for the evaluation of the clinical performance of PG10 scaffold.

CRediT authorship contribution statement

Cijun Shuai: Conceptualization, Methodology, Funding acquisition, Project administration. Feng Yang: Conceptualization, Methodology, Data curation, Writing – original draft. Yang Shuai: Formal analysis, Investigation. Shuping Peng: Validation, Investigation, Formal analysis. Shijie Chen: Methodology, Investigation, Validation. Youwen Deng: Methodology, Supervision. Pei Feng: Conceptualization, Methodology, Validation, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.